The Theory of Tariff Rate Quotas: An Application to the U.S. Sugar program using MONASH-USA

Full text

(1)CENTRE of POLICY STUDIESand the IMPACT PROJECT. Eleventh Floor Menzies Building Monash University Wellington Road CLAYTON Vic 3800 AUSTRALIA Telephone: (03) 9905 2398, (03) 9905 5112. from overseas: 61 3 9905 5112 or 61 3 9905 2398. Fax: (03) 9905 2426 e-mail web site. from overseas: 61 3 9905 2426 THE THEORY OF TARIFF RATE QUOTAS: AN APPLICATION TO THE U.S. SUGAR PROGRAM USING MONASH-USA by. ASHLEY WINSTON Centre of Policy Studies, Monash University. Preliminary Working Paper No. IP-83 July 2005. ISSN 1 031 9034. ISBN 0 7326 1557 7. The Centre of Policy Studies (COPS) is a research centre at Monash University devoted to quantitative analysis of issues relevant to Australian economic policy..


(3) 1. The Theory of Tariff Rate Quotas: an Application to the U.S. Sugar Program using MONASH-USA. Ashley Winston Centre of Policy Studies, Monash University. Abstract MONASH-USA (also known as USAGE-ITC) is a detailed dynamic general equilibrium model of the U.S. developed by the Centre of Policy Studies in collaboration with the U.S. International Trade Commission. This paper reports on the theoretical developments completed for a project intended to (a) add detail to the industry and commodity classification of the relevant sectors and (b) create a detailed modelling structure for the U.S. trade and industry-support policies that affect these sectors. A secondary theme of the research is the development and application of methods for modelling complementarity relationships in a large-scale general equilibrium framework. Data issues and simulation results are discussed in a second paper.. JEL Classifications: C61, C63, C68, D58, F13, F14, Q13, Q17, Q18. Keywords: international trade, agricultural policy, sugar, commercial policy, tariff-rate quotas, complementarity relationships, dynamic general equilibrium modelling.. * A shorter draft version of this paper was presented at the Conference on Global Economic Analysis held in Luebeck, Germany in June 2005. This paper contains a number of revisions. The views expressed in this paper (and the version presented in Luebeck) do not reflect those of the U.S. International Trade Commission or its Commissioners. Any errors are attributable only to the author..

(4) 2. Table of Contents THE THEORY OF TARIFF RATE QUOTAS: AN APPLICATION TO THE U.S. SUGAR PROGRAM USING MONASH-USA ............................................................................................................. 1 Abstract..................................................................................................................................................... 1 Table of Contents ...................................................................................................................................... 2 1 INTRODUCTION ........................................................................................................................................... 4 2 U.S. SUGAR AND CORN INDUSTRY POLICY................................................................................................. 5 2.1 The U.S. Sweetener Sector .................................................................................................................. 5 2.1.1 Sugar ........................................................................................................................................................... 5 2.1.2 Corn............................................................................................................................................................. 7. 2.2 The U.S. Sugar Support Program ....................................................................................................... 8 2.2.1 Background ................................................................................................................................................. 8 2.2.2 The Loan Scheme........................................................................................................................................ 9 2.2.3 The TRQ Program ..................................................................................................................................... 10. 3 CHANGING THE MONASH-USA THEORY TO MODEL TARIFF RATE QUOTAS.......................................... 11 3.1 Introduction ...................................................................................................................................... 11 3.2 The Microeconomic Foundations of Tariff Rate Quota Policies ...................................................... 11 3.3 Tariff Rate Quotas under the U.S. Support Programs ...................................................................... 17 3.3.1 Supply in the U.S. Sugar Industry ............................................................................................................. 17. 4 SIMULATING A CHANGE IN U.S. SUGAR POLICY ....................................................................................... 24 4.1 Introduction ...................................................................................................................................... 24 4.2 Adjustments to the Dimensions of Sets in MONASH-USA ................................................................ 24 4.2.1 Changes in Industry Sets ........................................................................................................................... 25 4.2.2 Changes in Commodity Sets ..................................................................................................................... 26 4.2.3 The Nesting Structure................................................................................................................................ 27. 4.3 A Generalised Complementarity Relationship.................................................................................. 37 4.4 The MONASH-USA Theory of the U.S. Sweetener Sector ................................................................ 40 4.4.1 Introduction ............................................................................................................................................... 40 4.4.2 Simulating the Sugar Program in MONASH-USA ................................................................................... 44. 4.5 Concluding Remarks......................................................................................................................... 59 BIBLIOGRAPHY ............................................................................................................................................ 60.

(5) 3.

(6) 4. 1 Introduction This paper addresses the nature of, and interactions between, U.S. sugar and sweetener industry support programs. A theoretical analysis is undertaken, and then used to inform a modelling structure for a large scale general equilibrium model of the United States. MONASH-USA, also known as USAGE-ITC, is a highly disaggregated, dynamic, computable general equilibrium (CGE) model of the United States developed at the Centre of Policy Studies (CoPS) in collaboration with the U.S. International Trade Commission (USITC).. The model. distinguishes over 500 industries and commodities, 51 regions (the 50 states plus the District of Columbia), over 700 occupational categories, multiple household categories and disaggregates international trade by the destinations and sources of exports and imports respectively. MONASHUSA is able to run dynamically with either backward-looking (recursive) or forward-looking expectations [or more detail on MONASH-style models, see Dixon and Rimmer (2002)] In early applications of MONASH at the USITC, concerns were raised about the model’s treatment of the sugar industry. The most significant areas of concern were (a) the level of detail captured in the industry and commodity classifications and (b) the theoretical structures underpinning the modelling of industry and trade policies. A program of development and model modification was undertaken to address these issues, the results of which are reported in this paper. In the first part of the paper, the structure of the U.S. sweetener sector is discussed, as well as the trade and domestic support policies applied to these sectors. Next, there is a theoretical analysis of the U.S. sugar program, placed in the context of conventional trade analysis under tariff rate quota (TRQ) regimes. This discussion suggests that care is needed in applying broadly defined theoretical norms to real-world policy regimes, and that regime-specific approaches are often required. Thirdly, a description is provided of the various adjustments made to the model’s database and dimensional structure. Some existing dimensions of the database are disaggregated to identify the various stages of sweetener-related agriculture and down-stream milling and refining, as well as the transformation possibilities available at the various stages of these processes (for example, cornbased stock feeds, corn starches and ethanol production as alternatives to wet corn-milling operations). Sugar imports into the U.S. are also disaggregated by source..

(7) 5 In the final section of the paper, a description is provided of the modelling techniques used to capture the various components of the trade and support policies directed at these industries.. 2 U.S. Sugar and Corn Industry Policy 2.1 The U.S. Sweetener Sector 2.1.1 Sugar The U.S. sugar industry is made up of the cane sugar and beet sugar industries. Unlike most sugar-producing nations, the U.S. has the geographical scale, climatic variation and domestic market size to support both types of sugar manufacturing solely for the purpose of supplying domestic consumption. The U.S. is not a direct exporter of sugar, although it does indirectly export sugar through trade in sugar-containing products. The cane sugar sector can be characterised as comprising three broad sub-sectors, 1. sugar cane farming, 2. sugar cane milling, and 3. cane sugar refining. Each of these sub-sectors produces a distinct product: 1. Sugar cane farming is an agricultural activity that produces sugar cane; 2. Sugar cane milling is an industrial activity that uses sugar cane as an input and produces raw cane sugar, as well as by-products such as molasses and sugar syrups; 3. Cane sugar refining takes raw cane sugar as an input and produces refined cane sugar. Sugar cane is grown in Florida, Hawaii, Louisiana and Texas. The U.S. sugar cane crop currently totals around 33 million metric tons per annum, of which about 90 percent is grown in Florida and Louisiana. The beet sugar sector can be characterised as involving two distinct activities, 1. sugar beet farming, and 2. beet sugar refining,.

(8) 6 and two distinct products, 1. sugar beets, and 2. refined beet sugar. Sugar beet farming regions are categorised by the USDA 1 according to a cost-of-production metric. In no particular order, these regions are 1. Great Lakes (Michigan, Ohio), 2. Upper Midwest (Minnesota and North Dakota), 3. Great Plains (Colorado, Montana, Nebraska, New Mexico, Texas, Wyoming) and 4. Far West (California, Idaho, Oregon, Washington State). U.S. sugar production in 2004 totalled approximately 7.83 million metric tons raw value (MTRV 2 ). This was comprised of 4.25 million MTRV of beet sugar and 3.58 million MTRV of cane sugar. The U.S. sugar industry typically ranks around 4th or 5th in the world by tonnage, depending on the season in question – for example, 2005 is expected to be an unusually low crop year due to the spate of hurricanes that hit Florida in 2004. The U.S. sugar industry generates a variety of products, but the bulk of its revenues are derived from sugar milling and refining. Other production possibilities for the mills and refineries, mostly by-products of sugar milling and refining, include bagasse, molasses, ethyl alcohol, rum, animal feeds, particle board and paper board. U.S. imports of all sugar products in 2004 totalled 1.56 million MTRV, comprised mostly of raw sugar imports under the TRQ and NAFTA programs of approximately 1.17 million MTRV, rendering the U.S. as the worlds 4th largest importer of raw sugar by tonnage for that year. The balance of the managed import volume is divided between refined sugar, sugar syrups and the indirect importation of sugar in sugar containing products. The management of sugar-containing. 1. 2. See, for example, any recent issue of ‘Sugar and Sweetener Situation and Outlook’, USDA, ERS.. A short ton is 2000 pounds, and a metric ton is 1000kg – so, 1 metric ton equals 1.102311 short tons. “Raw value” refers to a standard measure for sugar that takes into account the two levels of processing in cane sugar compared to the single level in beet sugar. One ton of refined sugar is equivalent to 1.07 tons of sugar, raw value..

(9) 7 product importing is also regulated by a TRQ policy, using a complicated set of formulae to calculate the sugar content of imported products, so as to apply quotas on the sugar content itself. The U.S. sources its imported sugar from 40 countries under the TRQ system, of which the largest suppliers in 2004 were the Dominican Republic, Brazil, the Philippines and Australia. Sugar cane must be milled soon after harvesting (usually within 72 hours), or the sugar yield decreases as the moist core dries. Similarly, beets must be processed very soon after harvesting in order to avoid a loss of extractable fructose. As a consequence, the agricultural and industrial components of the sugar industry (cane farms and sugar mills, and beet farms and beet-sugar refineries) must be located in relatively close proximity. As a consequence, regional sugar cane and sugar beet markets in the U.S. tend to face low competition from other regions, and cane and beet prices can vary substantially between sugar producing locales - even between districts within states. Refined cane and beet sugar are chemically indistinct forms of fructose, and in this sense enter the market as a single homogenous product. At the end of the production chain, therefore, refined sugars from the two sectors compete almost purely on price. The intermediate sugar milling sector of the cane sugar industry is the focus of the sugar program. The transportability of this product makes it most vulnerable to import competition. Raw sugar prices in the U.S. tend to be around 46 cents per kilogram, with the domestic support price under the agricultural loans program for sugar at around 39.6 cents per kilogram. The world sugar price, although historically volatile and variable through a wide range, is currently around 17 cents per kilogram.. 2.1.2 Corn The U.S. corn industry is a major agricultural producer and exporter for the US, with total production in 2004 in excess of 10 billion bushels (around 254 million metric tons 3 ). This level of production ranks the U.S. as far-and-away the largest producer of corn in the world, producing around 100 million tons more than the next largest producer (China). Around 50 million metric. 3. A bushel of yellow corn weighs 56 pounds..

(10) 8 tons of U.S. corn is exported annually to a range of countries, with the major destinations for U.S. corn being Japan, Mexico, Taiwan and Egypt. Most corn grown in the U.S. is yellow corn, a variety not typically grown for direct human consumption purposes (unlike sweet corn). Around 80 percent of the corn crop is used for stock feeds, with the rest going mainly to the production of wet and dry corn-milling products like cornbased sweeteners, corn starches and ethanol. In 2004, the use of corn in the U.S. wet corn-milling sector was divided between these three uses in the following proportions: 45.1 percent to sweeteners, 17.1 percent to starches and 37.8 percent to ethyl/methyl alcohols (mostly ethanol). When dry milling is included, ethyl/methyl alcohol production becomes the major milling product of corn, with total proportions being: 30.1 percent to sweeteners, 10.8 percent to starches, 51.7 percent to alcohols, and 7.4 percent to cereals. The dry corn-milling industry is the largest producer of ethanol in the U.S. – currently, only very small quantities of sugar based ethanol are produced or imported. The U.S. corn sweetener industry, a sub-sector of the wet corn-milling industry, produces mainly high fructose corn syrups (HFCS), glucose syrups and dextrose. Around 70 percent of cornsweetener industrial activity is accounted for by HFCS production. Corn sweetener production has expanded rapidly in recent years, and these products now account for around 56 percent of the total sweetener deliveries by-weight to U.S industrial demand and final consumption, having overtaken sugar as the major supplier of sweeteners in recent years. Notably, the U.S. beverage industry has almost entirely moved into using HFCS, while, to a lesser extent, manufacturers of baked goods like breads and cookies have increasingly switched to corn sweeteners for their baking properties.. 2.2 The U.S. Sugar Support Program 2.2.1 Background The current U.S. sugar program is built on the foundation of the Jones-Costigan Act of 1934. This act initiated a program of planned management for the U.S. domestic sugar market that (basically) involved (1) the forecasting of U.S. domestic demand on an annual basis, (2) the use of quotas as a means to formally allocate a proportion of the U.S. market to domestic and foreign suppliers, (3) the creation of marketing allotments (domestic production quotas) and (4) a formal program of acreage management designed to facilitate (3). This policy "template" has formed the basis of U.S. sugar policy ever since, with various revisions and modifications..

(11) 9 The Jones-Costigan Act was replaced with the Sugar Acts of 1937 and 1948, both of which carried the underlying principles of the original legislation right through to the early 1970’s. In 1974, the U.S. Congress repealed the Sugar Act of 1948, which lead almost immediately to large spikes in sugar production and a plummeting domestic sugar price. In 1977, Congress passed the Food and Agriculture Act, re-instigating a program of sugar industry support and supply controls, and introducing the first sugar loan program. The Agriculture Act of 1979 added a system of price supports to sugar processors, operating via the loan scheme, and authorising purchases of portions of the sugar supply by federal agencies, aimed at maintaining a stable producer-price and the management of U.S. consumer prices. In application, this was essentially a typical agricultural buffer-stock scheme aimed at price stability and support. The passing of the Food and Agriculture Act of 1981 saw the loan scheme modified to provide credit on a nonrecourse basis (that is, loans without penalties for delinquency), and was grandfathered essentially unchanged into the Farm Act of 1990. The 1996 Farm act again extended the sugar program, but limited the ability of federal agricultural authorities to intervene in markets and maintain prices above the loan rate (due largely to number of policy developments including U.S. commitments to the WTO and the NAFTA agreement). The 1996 Act also changed the nature of the loans from nonrecourse to recourse subject to the proviso that total sugar imports didn't exceed 1.336 million MTRV. The penalties instigated under the recourse loans scheme were 1 cent per pound for raw cane sugar and 1.072 cents per pound for refined sugar on any forfeiture (see below). Most recently, the 2001 Agricultural Appropriations Act eliminated the TRQ trigger for nonrecourse loans and removed all provision for loans to be granted on a recourse basis. The two most important components of the current U.S. sugar support program are the loan scheme and the TRQ program.. 2.2.2 The Loan Scheme Under the sugar industry support program, loans are taken out against raw sugar production at the “loan rate” (essentially a support price) with the sugar committed as collateral, ex ante, for a maximum term of 9 months. Initially, this loan rate was set at 16.75 cents per pound for raw sugar and 19.7 cents per pound for refined beet sugar, but has since been increased to 18 cents and 22.9 cents per pound, respectively. Loans must be repaid with any interest charges by September 30 in any given year. While cane mills are able to share the interest expense with their growers, beet refineries cannot and, so, must recover the entire interest expense from revenues generated by sugar.

(12) 10 sales – this in part explains the differences in the loan rates, along with the fact that beet sugar is typically sold at a 2 percent cash discount, as such. The Commodity Credit Corporation (CCC) administers the loans directly to sugar cane mills and beet sugar refineries, which qualify for the scheme by agreeing to offer sugar cane and sugar beet growers a minimum price for their produce of around 60 percent of the loan rate. As U.S. trade commitments under the WTO and NAFTA guidelines are autonomous to this decision making process, the marketing allotments effectively become the “free variable” in the hands of domestic authorities used as the policy tool, at least in the first instance, to limit the supply of sugar and maintain a target “minimum” price. The intent of the minimum price is to avoid sugar forfeitures, achieved by setting the price sufficiently high to cover the loan rate plus a small margin to account for mill-to-refinery transport costs and other distribution costs (plus, in some instances, location discounts imposed by refiners on mills) in the case of sugar mills, and exactly the loan rate in the case of beet refiners (as there is no second stage to processing and no additional costs are incurred by refiners post-production). Thus, the intention is make it more profitable for sugar mills and beet refineries to sell their output onto the market and repay the loans, rather than to forfeit the crop and retain the funds as a "compensation". Haley (2001) estimates that minimum prices required in 2000 to forestall sugar forfeitures were 19.86 cents per pound for raw cane sugar and 24.78 cents per pound for beet sugar, compared to the loan rates of 18 and 22.9 cents per pound respectively.. 2.2.3 The TRQ Program Under its obligations to the WTO set down in the Uruguay Round Agreement on Agriculture (URAA) of 1994, the U.S. agreed to import a minimum of 1.139 million MTRV of sugar in each marketing year (October to September). This is mostly comprised of raw sugar quotas, but does include a specific allocation of 24,251 MTRV of refined sugar. The allocation of TRQs to the 40 recipient countries occurs on a historical-allocation basis, calculated against the share of each country’s supply to the U.S. during the period 1975-81 when trade in sugar was deemed to be relatively free. In chapter 17 (heading 1701) of the U.S. Harmonized Tariff Schedule (HTS), the in-quota and over-quota rates applied to sugar, as negotiated under the URAA, are excise rates set (in general) at 1.4606 and 33.87 cents per kilogram respectively. The U.S. HTS allows for different rates for special cases, including zero in-quota tariff rates for nations annexed under the Caribbean Basin Economic Recovery Act (CBERA) and those covered by the Generalized System of Preferences.

(13) 11 (GSP). These exemptions, in practice, have lead to a situation where most of the 40 nations holding TRQ rights are supplying sugar to the U.S. market free of U.S. duties on in-quota amounts. Additional quotas over-and-above any allotments under the TRQ program are allocated to Canada and Mexico under NAFTA. Canada supplies small quantities of beet sugar to the U.S. market, while Mexico supplies both refined and raw cane sugar. These quotas are discussed further below.. 3 Changing the MONASH-USA Theory to Model Tariff Rate Quotas 3.1 Introduction The microeconomic foundations of TRQ policies have been well-covered in the literature. However, this literature typically relies on simple, generic frameworks that suppress important details like those evident in the United States sweetener sector policies. Furthermore, generic approaches don't lend themselves readily to welfare analysis. This section addresses and expands the standard approach, with particular reference to the U.S. sugar policy.. 3.2 The Microeconomic Foundations of Tariff Rate Quota Policies In application to agricultural policies, imposed on industries with annual production/cropping cycles and significant lags between production decisions and revenue receipts, industry supply response and the implications of the resulting shifts in market power are important factors. The nature of U.S. sugar supply in a given period differs from that of a ‘generic’ commodity. A series of diagrammatical analyses is developed below that begin with, and then extend and refine, the conventional approach to TRQs to incorporate the detail of U.S. sugar programs. In the diagrams discussed below, applied tariff rates are ad-valorem equivalent unless otherwise noted. Denote the in-quota and over-quota ad-valorem equivalent (see below) tariff rates by ti and to respectively; the quota by Q; imports by M; the world price (in U.S. dollars) by Pw; the U.S. customs-insurance-freight (CIF) price by PCIF; the U.S. landed-duty-paid (LDP) price by PLDP or PUS; and the rent per unit (or premium) generated by the quantitative constraint at the sugar quota boundary by φ ..

(14) 12 Denoting the excise rate (using the in-quota value for this example) by tie , the ad-valorem equivalent is determined by recognising that any level of revenue generated by an excise tax has an equivalent value-based rate implied by the equality tie M = ti PCIF M ,. which implies. ti =. tie . PCIF. (1). When the ‘powers’ of the tariffs (i.e. the ad-valorem tariff rate plus 1) are applied in this paper, this allows the multiplier to be inserted either below or above the rent rectangle in the diagrams when the appropriate price variable - PCIF - is applied in the denominator. The CIF price is equal to the world price plus rents, or from an alternative perspective, the LDP price minus the tariff. Diagrammatically, the tariff-revenue rectangle remains a constant absolute size as either of these prices varies, which is particularly useful in the case of the rent-inclusive CIF price when analysing on-quota imports, or when plotting multiple demand schedules on a single diagram..

(15) 13 Figure 1. $US. PUS3 = Pw(1+ti)(1+ φˆ ) = Pw(1+to). s. l. t. m. D3 PUS2 = Pw(1+ti)(1+ φ1 ). PUS1 = Pw(1+ti) Pw. g. h. e. f. b. c. i j. D2. n. D1 a. d M1. k M2. o M3 Imports. Hereafter, the variable φˆ denotes the maximum value of φ . Representing a fairly simple approach to TRQ analysis, figure 1 highlights the important underlying features of this type of trade policy. Three different levels of demand for sugar are plotted in figure 1, given the values of the two ad-valorem tariff rates and the sugar quota, to represent in-quota (D1), on-quota (D2) and over-. quota (D3) import volumes. Demand schedule D1 captures sugar imports occurring at an in-quota level M1, attracting the inquota tariff and generating zero rents as the quantity constraint imposed by the quota is not binding. The US CIF price is equal to the U.S-dollar world price, the total CIF value of imports is abcd, the LDP value is aefd, and tariff revenue is represented by area befc. The LDP price – the basic price to U.S. agents represented along schedule D1 - is PUS1 in this case. Demand schedule D2 generates an on-quota level of imports, M2, equal to the quota Q. As the quantity constraint is binding, a rent accrues to the foreign exporter and the CIF price rises accordingly. Under these circumstances, the CIF value of imports increases to aghk minus tariff revenue eijb, and now involving a rent captured by area ghie. These values are generated by the onquota level of imports Q, the LDP price PUS2 (inclusive of a rent and tariff wedge) and a CIF price.

(16) 14 equal to PUS2/(1+ti).. As demand expands, the tariff revenue remains constant and the rents. generated by the quantitative restriction expand proportionately with the CIF price - that is, not proportionately with the LDP price as the tariff is an excise tax. In this example, excess demand at PUS1 is generating a per-unit rent of φ1 , which is assumed to accrue to the foreign supplier who holds the quota “licence”, as is the case in the U.S. sugar program. In the case of demand schedule D3, the import volume increases to an over-quota level M3, the LDP price increases to PUS3, the CIF price at the margin falls to PUS3/(1+to) - again equal to Pw – on the over-quota import quantity (M3-Q). Tariff revenue now comes in two parts – a fixed component given by the in-quota tariff rate ti and the quota Q (area eijb), and a second component determined by the over-quota tariff rate to and the level of over-quota imports (M3-Q) (area lmnj). As demand drives the LDP price to reach PUS3, the per-unit rent reaches a maximum value of φˆ , but this is only applicable to sugar imported under the quota. Thus, as the volume of imported sugar goes overquota, the total value of rents has reached a maximum also, given by area slie. How complete is this analysis? Ignore the specific characteristics of sugar production and assume a generic commodity. If the imported good is determined to be readily substitutable for a U.S. variety, the three demand schedules in figure 1 must be interpreted as excess demand schedules. As such, a diagram like figure 1 cannot be used for analysing the affect of trade policies on domestic economic welfare under such circumstances, as it entirely ignores the implications of policy for (a) domestic production of the substitute and (b) overall consumption behaviour. It seems unlikely that a trade policy would be applied to a good that had no domestic substitutes in need of protection and thus, logically, the diagram is incomplete. Figure 2 below addresses some of these issues..

(17) 15 Figure 2. D3. $US. F G. D2. Q. D1 PUS3 = Pw(1+ti)(1+ φˆ ) = Pw(1+to) o. t. h. i. j. C a. b. c. PUS2 = Pw(1+ti) (1+ φ1 ) PUS1 = Pw(1+ti) Pw. y. x. w. H. u k d E. l. m. n. e. f. g. M3-Q. A S. d. M3. D3. SdD2 SdD1. M2=Q M1 S1. S2. S3 Total Quantity. Figure 2 is the underlying market diagram that generates figure 1, drawn using the assumption that a close domestic substitute for the imported good exists. The demand curves in figure 2 plot the underlying demand schedules that (along with the domestic supply schedule AF) generate the excess demand curves plotted in figure 1. It is helpful in reconciling these two diagrams to recognise the following: (1) the curve Sd in figure 2 is equivalent to the vertical axis plotted in figure 1; and (2) the curve hlxH in figure 2 is equivalent to curve bjlt in figure 1. The “total” supply schedule – the horizontal sum of domestic and foreign supply at every price - is given by line ACg in the absence of any trade policies. With the TRQ system in place, line AhlxH describes the supply curve, with segment hlxH being equivalent to the importsupply curve plotted in figure 1. As a demand schedule is moved to the right and passes through point l, imports are restricted by the quota and domestic production responds to the increase in price as excess demand materialises. The increase in the domestic price at the quota boundary is also passed to the foreign supplier of the import quota as a rent, and the supply curve changes as the line segment hl slides up the underlying domestic supply curve AF. Once demand is sufficient to generate the LDP price PUS3 at point x, buyers are able to purchase as many units of the good as desired from foreign suppliers at this price, equal to the world price Pw plus the over-quota tariff..

(18) 16 As such, a further positive shock to demand must be met by an increase in imports as domestic production has reached a ‘maximum’ at that price. At point x and beyond, the line segment that previously represented the quota ceases to “slide” along AF and, instead, attaches to point w and extends out to the right of the diagram. Figure 2 lacks the simplicity of figure 1, but therefore includes detail that is subsumed in a figure 1 style diagram that focuses on excess (domestic) demand and imports. A comprehensive welfare analysis using a diagram like figure 2 will be conducted later in the paper. A complete welfare analysis is not possible in figure 1 unless, as mentioned, there is no substitutability between the imported good and a domestic commodity, in which case there would be no “figure 2”. Of most importance, in the event that substitutability does exist, is that the domestic supply response is included so that it becomes possible to examine how consumer and (total) producer surplus areas are redistributed by changes in policy variables. For now, a brief interpretation of figure 2 that enables comparisons with figure 1 is provided. With demand schedule D1, total supply is found at S1, comprised of domestic production SdD1 and in-quota imports M1, and generating tariff revenue ahkd. As this level of imports lies below the quota (hl) by the amount kl, there are no rents generated for the foreign exporter. This was assumed implicitly in figure 1 also. It should be noted that this is not always true – the discussion of the U.S. sugar program below will highlight the circumstances under which this assumption does not hold. For now, keep in mind that the quotas in figure 2 are not interpreted as import commitments by the importer, and that domestic supply is responsive to current market conditions. Under demand conditions represented by curve D2, imports have reached the quota boundary and rents are being generated. In this case, the equilibrium levels of total demand and supply occur at S2, where supply is composed of domestic production SdD2 and imports M2, equal to the quota (also represented by line segment ou in this case). Tariff revenue is given by area bime, and rents by ioum. Compared to the previous case, this higher level of demand stimulates an increase in domestic output by hi, and an increase in (domestic) producer surplus of area hio - this is absent from figure 1, as is the change in consumer surplus due to increased consumption of the domestic variety and the transfer of surplus from consumers to domestic producers caused by the price increase (both of which we ignore for now in figure 2). Demand curve D3 generates an equilibrium total level of demand and supply at S3, comprised of domestic output SdD3, and imports M3, equal to the quota plus over-quota imports (M3-Q) along line.

(19) 17 segment xy. Total tariff revenue is now a function of two components, (1) imports at the in-quota tariff rate generating revenue cjnf on quantity wx (i.e. the quota) plus (2) imports at the over-quota tariff rate generating revenue fxyg on quantity xy. Once the LDP price has reached PUS3 and the level of imports has exceeded the quota, the total value of rents reaches a maximum value of rmaxQ, captured in figure 2 by area jwxn. Compared to the outcome generated by demand schedule D1, curve D3 has generated an increase the domestic price from PUS1 to PUS3, and stimulated (1) an increase in domestic output hj and (2) domestic producer surplus hjw, larger than the case of curve D2 by ij and ijwo respectively.. 3.3 Tariff Rate Quotas under the U.S. Support Programs 3.3.1 Supply in the U.S. Sugar Industry The ‘current’ spot price and quantity are measured along the vertical and horizontal axes, respectively, of figures 1 and 2. The implication is that an immediate supply response is possible from all sectors of the market in all situations where the quota is not binding. This assumption is not defendable in the U.S. sugar market. At the time that sugar enters the market and prices are determined in any given year, that year's level of sugar production is highly inelastic with regard to that year’s sugar price. As discussed in section 2.2 above, the U.S. sugar program involves domestic production quotas – known as “marketing allotments” – and a commodity loan scheme in order to maintain an implicit price floor of approximately 39.6 cents per kilogram (plus a small margin, as discussed above). The program aims to avoid loan default and sugar forfeitures. Given this aim, and in light of the empirical evidence found in historical stock-accumulation data, it seems reasonable to assume that sugar stocks do not systematically accumulate. In administering the program and managing market supply in pursuit of the price floor (i.e. the ‘loan rate’), a forecast is generated for the U.S. total demand for sugar at about this price floor plus a margin for error. The import commitments under WTO obligations are subtracted from this forecast, and the residual determines the level of domestic output required to bring total sugar supply to the level required to achieve and maintain the price floor. The domestic component of total U.S. sugar supply is allocated to the domestic industry in the form of production quotas (i.e. ‘marketing allotments’). As is the case for many tradable agricultural commodities, sugar production is characterised by long lags between decision making and outcomes. Production decisions are made on the basis of.

(20) 18 expected prices well before the output is released onto the market and revenue streams are generated. Therefore, other than for the real possibility of over-quota imports from Mexico at the lower NAFTA over-quota tariff rate, sugar supply above the price floor at the time of harvesting is highly inelastic. Supply is also very inelastic below the price floor at this time, as costs are sunk at the time of harvesting and the beginning of the marketing cycle. If the variable costs of growers and millers cannot be covered in the event of extremely low prices, crops may be abandoned and mills run at less than full capacity utilisation or, indeed, left idle – the equivalent of the “shutdown” point of undergraduate microeconomics classes. Some limited supply response is afforded the authorities through the loans scheme, as it is within their power to withhold forfeited sugar from the market if demand forecasts have been too optimistic.. A further small degree of supply. responsiveness flows from any sugar stock held over from previous years, but the change in these stocks is small on average, implying that their role in the behavioural outcomes that determine the total sugar supply is marginal.. Figure 3. Domestic Supply. NAFTA Imports $US. NAFTA Capacity NonNAFTA Imports. Pw(1+to) PN(1+tN,o). PS Marketing Allotments NAFTA PSD TRQ Pw. Sugar. Figure 3 plots the components of total U.S. sugar supply at the time that the marketing cycle begins. As the U.S. WTO commitments are binding at this time, and as the sugarcane has been harvested and sugar mills are processing sugar, domestic supply is rendered highly inelastic at a.

(21) 19 level equal to the sum of the marketing allotments. As mentioned, very low prices may stimulate crop abandonment and create excess capacity in sugar mills, and therefore - at the limit - supply could potentially fall to zero. This is a situation unlikely to eventuate in reality, as it would require the convergence of large and unlikely errors in the expectations of cane growers, sugar processors and the federal authorities who set the marketing allotments. The domestic supply schedule in figure 3 is drawn to show that U.S. sugar production costs are higher than both the world price (inclusive of the in-quota tariff) and the NAFTA duty-free price.. The diagram defines a. “shutdown” price PSD for domestic production – at prices slightly higher than this, profits are likely to be negative but some fixed costs can be covered after variable costs have been met, while below it a cane grower or milling firm cannot even cover variable costs. The TRQ curve plots the U.S. WTO commitments for imported sugar. This amount must be imported under these WTO commitments, and so is autonomous with respect to market spot prices during the marketing cycle. At a domestic price below the tariff inclusive world price, no sugar would be offered for sale by foreign exporters, although, again, WTO commitments in effect oblige the U.S. to purchase this amount at a price no less than the world price. The NAFTA supply schedule is also drawn as relatively inelastic.. Under the NAFTA. agreement, Mexico receives an additional quota allocation over-and-above the U.S. WTO obligations. This amount is calculated, on a year-by-year basis, as a function of the estimated Mexican sugar surplus (of production over consumption) up to a maximum amount of 250 thousand metric tons. In practice, the allocation is usually set at a level far below this. The NAFTA curve becomes perfectly elastic at the Mexican over-quota tariff rate (reflecting that Mexican sugar production costs are significantly less than this implied price) and again becomes inelastic at the point that Mexican supply capacity is reached for a given season. The Mexican quota can be filled with any combination of raw and refined sugar: this sugar enters the U.S. market duty-free within the quota allocation and at a relatively low over-quota tariff rate of around 10 cents per kilogram in the event that the quota allocation is exceeded. In most recent years, Mexico has imported more sugar at world prices than it has exported to the U.S. at the U.S. price, effectively making a healthy arbitrage profit on its quota allocation. As the quotas and duties are to be removed in 2008 under NAFTA, this highlights an issue of concern for U.S. trade policy makers. Rules of origin apply currently to combat this type of behaviour, but these rules do not effectively limit the ability of a country to export its own product while consuming imported sugar..

(22) 20 Combining these functions generates the effective U.S. supply curve for sugar in any given year, relating current prices and production, and is plotted in figure 4 below. Note that the two axes in each of figures 3 and 4 are not drawn to scale.. Figure 4 $US D Pw(1+to) PN(1+tN,o). PUS. PS. Pw(1+ti) Pw. Sugar. Figure 4 indicates that the supply of sugar in the U.S. during the marketing cycle is highly inelastic throughout a wide range of prices, and only involves some price-responsiveness at very low prices (due to crop abandonment, etc) and at prices sufficiently high to stimulate over-quota imports from Mexico or the rest of the world. This supply curve indicates the opportunity cost of sugar to the various supply-side agents only once the commitment has been made to supply the various quantities, and in this sense is not truly a “supply” function, telling us little or nothing about the underlying resource allocation decisions. Due to these limitations, it is not possible to conduct a complete behavioural analysis with figure 4. As our interest in economic policy analysis rests on concepts such as allocative-efficiency, redistributive effects, and the behavioural and welfare outcomes of production decisions in the sugar industry (both domestic and foreign), we need to relate the expected market price of sugar to the underlying production functions and the opportunity cost of resources allocated to this use..

(23) 21 Figure 5. $US. y z. MCUS. Total Quotas. D. w A E. B. SF1(1+t). C. PS. u v. t. w. x. q. p. o j. i. r m k. n l h. c. d. g. f. b. a. S1. Marketing Allotments NAFTA. ROW. SF1. s. SF2. e. S2 Total Quantity. Figure 5 takes some components from figures 2 and 4, along with some pertinent characteristics of the U.S. sugar industry, and combines them to generate a more comprehensive diagram representative of the U.S. market for raw sugar.. Domestic production is limited by quotas. (marketing allotments) to ab, NAFTA (Mexico) is allocated a duty free quota bc, and the “rest-ofworld” (ROW) is allocated quotas summing cd 4 . ROW imports are subject to an in-quota ad valorem equivalent tariff rate of mk/kc, and an over-quota rate of Al/kc. NAFTA raw sugar can be imported over-quota at an ad valorem equivalent rate of Cl/kc, up to a maximum amount (determined by net capacity in any given year) of BE. The diagram assumes that NAFTA and ROW imports are (on average) produced according to the same production function – this is clearly not a defendable assumption, but a relaxation of the assumption changes none of the qualitative. 4. Note that these quantities are not drawn to scale..

(24) 22 outcomes of the analysis represented here. The foreign supply curve SF1 plots a price-responsive supply of foreign sugar (including NAFTA sugar), reflecting (a) that the U.S. is a major importer of raw sugar and (b) the global capacity constraint in agricultural production in any given year. In order to divert sugar from other buyers on the world market, the U.S. would need to offer a higher price. Foreign supply curve SF2 is simply a horizontal shift of SF1 by the amount of the domestic marketing allotments, and indicates that a CIF price of ld would be charged on the quota volume in the absence of the domestic support program. Domestic supply is plotted by curves iqv and xw (involving a discontinuity across vx), produced at a marginal cost determined by the curve tvxMCUS (also with a discontinuity along vx). The diagram is drawn to represent a situation of perfect foresight – the marketing allotments have successfully filled the gap between domestic raw sugar demand and the U.S. import obligations such that the cost of the marginal unit of domestic production is just covered, and there are no incentives for domestic producers to expand output (if they could). This ignores the margin required to cover additional expenses that was discussed above, and so assumes that the loan rate is equal to the trigger price for forfeitures. Figure 5 suggests that no domestic raw sugar would produced under a free trade regime. Equilibrium output and consumption is determined at S2 at a price of se, generating total expenditure apse. Total welfare is given by area ysf, of which consumer surplus xsp accrues to domestic agents and producer surplus psf to foreign producers.. This outcome has obvious. consequences for ‘upstream’ cane growers that are not captured in this diagram - however, offsetting this to some degree, it is likely that cheaper raw sugar would induce a demand response from downstream users, and sugar refineries and other end users of raw and refined sugar should see an expansion in their activity in response to lower input costs. This is an example of the type of general equilibrium linkage that MONASH-USA is designed to account for. Under the current U.S. support program, the supply of sugar S1 at the support price PS generates total expenditure auxd, a change of drse minus puxr. Which of these expenditure-change areas is the larger depends on the arc elasticity of demand around line segment xs: however, based on the econometric evidence of relatively inelastic raw sugar demand in the U.S, this is likely to stimulate an increase in expenditure 5 . Total economic welfare derived from production and consumption in. 5. See, for example, Haley (1998)..

(25) 23 this market is represented by area yxlgvt, comprising consumer surplus yxu to domestic consumers, uvt in producer surplus to domestic producers, tariff revenue kmnl accruing to the domestic government, and some foreign surplus areas. These foreign surplus areas can be divided into mwxn in rents to ROW producers, jvwk in rents to NAFTA producers, and producer surplus area jlg going to both in proportions determined by their share in the total quota, bc/bd for NAFTA and cd/bc for ROW (assuming identical production functions). The outcome for foreign producers depends on the relative sizes of (1) the loss in producer surplus from the decrease in volume and (2) the various rent rectangles, while for the U.S. the outcome is an unambiguous loss of total welfare equal to ptvxs minus the tariff revenue kmnl. The overall welfare loss to the U.S. is a function of the deadweight production-side loss ptvq (due to the increased cost of domestic production over foreign acquisition) a deadweight demand-side loss equal to rxs (due to the increase in price and reduction in consumption) and a transfer of consumer surplus to foreigners as rents equal to qvxr. The efficacy of a TRQ and support price program can be addressed by a comparison with a tariff-equivalent policy that involves an ad valorem tariff sufficient to generate the same price and total volume outcomes. This has been made possible on figure 5 with the addition of a dashed foreign supply curve SF1(1+t). The equivalent-tariff case enables the domestic government to internalise all rents available under the TRQ policy in the form of tariff revenue, and in doing so (as the diagram is drawn in this case) stimulate no domestic production increase and associated production-side deadweight losses. Of course, underlying this analysis is the assumption that factor markets will clear and that the domestic resources freed up by import substitution are utilised in more valuable alternative uses. In the presence of wage floors and market frictions (e.g. search costs, information asymmetries, etc) this assumption might be questioned, but it is a well-accepted tenet of economic theory and empirics that trade policy is a poor substitute for labour market programs in pursuing full employment over the medium to long term. Figure 5 generates some outcomes from the TRQ policy that are not typical of this type of trade analysis. Of significant influence in this matter is that sufficient quantities of foreign-sourced sugar to clear the U.S. market can be supplied at a lower cost than the first unit of domestic production, implying that the domestic industry ceases to exist without trade restrictions. However, this is partly generated by some simplifications, perhaps most importantly the assumption of low price elasticities for import supply across the relevant range. The U.S. is committed to around 1.5 million metric tons of sugar imports out of total global trade in sugar of around 35 million metric tons. This is a relatively small, but not insignificant, part of the total. In the event that the U.S. chose to increase sugar imports significantly, it is likely that capacity constraints (generated by the nature of.

(26) 24 sugar production) applying to global supply in a given year would see the world spot-price move to divert significant quantities of sugar from other destinations. It is clear, in any case, that the U.S. cane growers and raw sugar milling installations would be adversely affected by a free and open sugar market in the United States. This is not the end of the story for total U.S. economic welfare however, as general equilibrium analyses of such trade policies tend to show.. 4 Simulating a Change in U.S. Sugar Policy 4.1 Introduction The first part of this section provides a brief overview of changes to the structure and database of MONASH-USA. A number of new industries and commodities were added to the model, requiring that some existing industries and commodities were split and redefined. These new components also required that appropriate nesting structures be determined, but this issue is left to the discussion of the model’s theory below. To simulate the U.S. sugar program realistically, it is necessary to impose some inequality constraints (for the TRQ quotas) and apply some discontinuous (and thus non-differentiable) functions (for tariff revenue and quota rents).. These relationships can be represented by. complementarity functions that relate the values of two or more variables in various states. Generally speaking, applications of general equilibrium models to analyse quantity restrictions on trade such as quotas and TRQs have utilised implicit price premiums as proxies for the required inequality constraints. The GEMPACK software developed at the Centre of Policy Studies by Pearson et al, designed to solve very large systems of non-linear equations in general equilibrium models, has recently been enhanced with dedicated sub-routines that enable inequality constraints and discontinuous functions, like those required for complementarities, to be modelled in a theoretically rigorous way. In the second part of this section, the application of these sub-routines and some numerical methods to the modelling of the U.S. sugar market is described.. 4.2 Adjustments to the Dimensions of Sets in MONASH-USA To begin this exposition of the adjustment to MONASH-USA, it is useful to examine the changes that were made to the industry, commodity and regional dimensions of the model.. After discussions with various industry groups/lobbyists and several industry.

(27) 25 analysts from the Office of Industries at the U.S. International Trade Commission, it was decided that the following adjustments would be made to the models dimensional structure.. 4.2.1 Changes in Industry Sets In MONASH-USA prior to this project, the industry dimension contained 513 elements, of which the relevant sectors were: •. I7 Feed Grains. An industry mainly engaged in growing corn, barley, sorghum, and oat crops;. •. I13 Sugar Crops. An aggregation of the sugar cane and sugar beet growing industries;. •. I75 Wet Corn Milling. An industry that is defined by its processes, but which mainly produces high-fructose corn syrups (HFCS), other corn sweeteners (glucose syrups and dextrose), corn starches, and corn (ethyl) alcohols;. •. I79 Sugar. An aggregation of the cane and beet sugar processing industries, producing raw sugar (and by-products) at sugar mills, and refined sugar (and by-products) at sugar refineries.. These four industries have now been expanded and redefined to create eight sugar and corn related industries: •. I7 Corn Crops. An industry engaged in the growing of corn crops, which comprised around 80% of the output by volume of the ‘Feed Grains’ industry previously;. •. I8 Other Feed Grains. Effectively the residual of the old ‘Feed Grains’ industry, engaged mainly in growing barley, sorghum, and oat crops;. •. I14 Sugar Cane. The farming of sugar cane crops;. •. I15 Sugar Beets. The farming of sugar beet crops;. •. I77 Wet Corn Milling. Essentially as defined before, but now with explicit account taken on the output side for multiple products.

(28) 26 through transformation possibilities imposed by CET transformation functions in the production nests; •. I81 Raw Sugar Milling. An industry involved mainly in milling sugar cane and producing raw cane sugar, but also by-products such as molasses, brown sugar, and sugar syrups;. •. I82 Cane Sugar Refining. An industry engaged in refining raw cane sugar into refined sugar;. •. I83 Beet Sugar Refining. Like I82, but involves the refining of sugar beets into refined sugar.. These modifications expand the MONASH-USA industry dimension to contain 517 elements.. 4.2.2 Changes in Commodity Sets A similar process is followed for the commodity dimension of MONASH-USA. Of course, the industry and commodity dimensions are closely tied together. Before modification, MONASHUSA accounted for 503 commodities, including: •. C7 Feed Grains. Corn, barley, sorghum, and oats, produced by the feed grains industry;. •. C13 Sugar crops. Sugar cane and sugar beets, produced by the sugar crops industry;. •. C74 Wet corn milling. An aggregation of HFCS, other corn sweeteners, corn starches, and corn (ethyl) alcohols, produced by the wet milling industry;. •. C78 Sugar. Essentially a composite commodity of all output from the sugar industry other than the agricultural component of production.. These 4 commodities were split into nine new commodities: •. C7 Corn. Only corn, to be produced by the new corn industry;. •. C8 Other feed grains. Barley, sorghum, and oats, to be produced by the new ‘other feed grains’ industry;. •. C14 Sugar cane. Only sugar cane, grown by the sugar cane industry;.

(29) 27 •. C15 Sugar beets. Only Sugar beets, produced by the sugar beet industry;. •. C76 Wet milled sweeteners. Corn based sweeteners (HFCS, dextrose, glucose syrups), produced by the wet milling industry;. •. C77 Wet milled other. Corn starches, ethanol, etc produced by the wet milling industry;. •. C81 Raw cane sugar. Raw cane sugar and by products, produced by the raw sugar milling industry. •. C82 Refined cane sugar. Refined cane sugar and by-products, produced by the cane sugar refining industry;. •. C83 Refined beet sugar. Refined beet sugar and by-products, produced by the beet refining industry.. The modified MONASH-USA model now contains a total of 508 commodities.. 4.2.3 The Nesting Structure The following diagrams capture how these new industries and commodities are related in the model’s theory. Firstly, figure 6 provides an overview of the sugar industry in MONASH-USA. abbreviations dom and imp referred to the domestic and imported varieties respectively.. The.

(30) 28 Figure 6 FINAL USES. FINAL USES. C82 Refined Cane Sugar. C83 Refined Beet Sugar. CES NEST (Armington). CES NEST (Armington). C82 Refined Cane Sugar dom. C82 Refined cane Sugar imp. C83 Refined Beet Sugar dom. CRESH NEST. CRESH NEST. Multi-source by TRQ allocation. OTHER USES. C83 Refined Beet Sugar imp. I82 Cane Sugar Refining. I83 Beet Sugar Refining. Multi-source by TRQ allocation. C81 Raw Cane Sugar. Processing. CES NEST (Armington) C81 Raw Cane Sugar dom. C81 Raw Cane Sugar imp. I81 Raw Sugar Milling. CRESH NEST Multi-source by TRQ allocation. C14 Sugar Cane. C15 Sugar beets. I14 Sugar cane. I15 Sugar beets. Primary Production.

(31) 29 At the primary production level for cane sugar, the sugar cane growers produce the cane and sell it into the intermediate stage of the sugar processing industry, cane milling. These mills crush and extract the cane juice from the cane, which is boiled and clarified to produce molasses and syrup. These are then put through a crystallization process that creates the sucrose crystals - raw cane sugar crystals. Sugar cane is not imported, as it does not travel well – after 72 hours, sugar cane begins to dry and lose its extractable fructose content. On the other hand, raw cane sugar is a highly transportable, homogenous bulk commodity, and faces significant import competition. The main intention of the U.S. sugar program is the maintenance of the domestic raw cane sugar industry (and thus also supports the cane growers). Imports of raw sugar are sourced from around 40 countries under the U.S. TRQ program.. These are firstly combined in a CRESH nest. parameterised with relatively high substitution elasticities. These parameter values are set high in order (a) to ensure that different import regions are treated as supplying a relatively homogeneous and highly substitutable product and (b) to generate the right type of behaviour in the complementarity statements used to model the TRQ program (this is explained in more detail below). The composite imported sugar product that results from this nesting is then combined with the domestic variety in an Armington (CES) nest, again with high substitution elasticities to ensure that the two goods compete almost entirely on the basis of price. The raw cane sugar composite is sold on to the cane sugar refining industry (which produces refined sugar), to other industries (some food industries, for instance) and to final demand. Sugar refining is a process involving the addition of water to the raw sugar crystals to create a liquid, followed by repeated heating and clarifying and a crystallization process. Small quantities of refined sugar enter the U.S. under the NAFTA and TRQ programs, and these are nested in a CRESH regional nest, and then in an Armington nest with the domestic variety. At this point it is clear, even from a qualitative perspective, that the fortunes of the refined sugar industry and downstream raw and refined sugar users are not served by the sugar program – from a refiner's or sugar-user's perspective, it simply raises the cost of an input. The lack of complaint from sugar refiners on this issue may hint at the existence of some interesting underlying industry ownership structures, or perhaps at some less formal structures involving effective vertical integration and cost shifting/transfer pricing behaviour. Sugar beets are produced by the sugar beet growers, and sold directly to the beet sugar refineries – there is no intermediate stage for beet sugar processing. Beets are washed, sliced into ‘chips’ and.

(32) 30 soaked in water, after which the pulp is leached from the juice. The juice passes through further processing steps, including clarifying and thickening stages, and is finally subject to a crystallization process identical to that used for cane sugar. Small quantities of beet sugar are allowed to enter the U.S under the NAFTA and TRQ programs, but mainly from Canada. These import regions are nested via a CRESH regional nest into an imported composite, and then with the domestic variety in an Armington nest – both again utilising relatively high substitution elasticities. Figure 7 outlines the structure of the corn industries in MONASH-USA..

(33) 31 Figure 7. FINAL USES. C76 Corn Sweeteners. CES NEST (Armington). C76 Corn Sweeteners dom. C77 WetMilling other. C76 Corn Sweeteners imp. Processing CET NEST OTHER USES (mainly stock feeds). I77 Wet Corn Milling. C7 Corn. I7 Corn Crops. Primary Production. The new corn cropping industry in the model grows and harvests corn. The bulk of the U.S. yellow corn crop which is sold to animal feed-stock uses, while a relatively small proportion is sold to the wet milling industry. Wet corn milling plants use the corn to produce a mix of three products – corn sweeteners, corn starches and ethanol – through a CET (constant elasticity of transformation) function..

(34) 32 In recent years, there has been an evident willingness in the wet milling industry to switch production between products based on relative output prices.. The USDA’s ‘Feed Outlook’. publications often refer to the fortunes of the corn producers with reference to conditions in the ethanol and corn sweetener markets. This is the rationale for using the multi-product structure for the wet-milling industry, which allows the industry to determine its output mix based on the market prices of the goods it produces. The wet-milling of corn is process distinguished from dry-milling by the soaking of the corn husks in water prior to milling. Corn is steeped and then is sent through a series of processes (grinding, screening, starch-gluten separation, starch conversion and fermentation) that lead (at different stages of this process) to feed products, starches, sweeteners and alcohols. Currently the model takes account of corn sweetener imports, which are nested Armington-style with the domestic variety before being sold to various end-users. Ironically, the fortunes of firms involved in corn sweetener production are partly dependent on the existence of the sugar program even though they compete with sugar producers in the market for sweeteners. Unit costs for corn sweetener production are falling, but are still significantly higher than the world price of sugar. Thus, although the competition is fierce between the two sweetener products, it is likely that the corn-sweetener industry would suffer substantially if the sugar program was abandoned, as it would be likely to lose much of its market to imported sugar. The corn industry is at pains to point out that the two goods are not close substitutes, in part because they say that industry has changed plant and processes to use HFCS (mainly) and is not able to directly substitute sugar. This is only credible in the very short run, however, as the recent rapid move into HFCS and away from sugar illustrates – there is no reason to assume that a significant fall in the sugar price could not stimulate a reversal of this change in technique for food production. It is clear that producers will be (and have been in the past) willing to bear the large fixed costs of technique adjustment if the return on the alternative plant over time warrants it. This, in fact, is the crux of inter-temporal optimisation behaviour that underpins neo-classical investment theory. Another argument used by the corn lobby, and heard by the author from several sources, rests on the claim that variable costs in the wet-milling industry are very low (around 8 cents per pound). The argument continues to suggest it is unlikely that the domestic CIF price for sugar would ever fall to that level, and thus it is unlikely that end-users of sweeteners will ever switch back into sugar and away from corn sweeteners in response to price differentials. This argument shows a partial understanding of microeconomic theory and industrial structure, but not enough of an.

(35) 33 understanding to make the argument credible. Elementary microeconomic theory shows that output decisions in the short run are made according to the variable cost and revenue functions. This is the essence of the industry's own argument - as long as the variable cost of HFCS production is lower than the price of imported sugar, there will be no substitution away from HFCS. However, while this is potentially true in the short run, the large fixed costs evident in the industry’s cost structure renders it likely that firms are making losses on HFCS production on prices far higher than 8 cents per pound. In the very short run in the presence of effectively fixed primary factor usage (capital and land), it might well be true that firms continue to produce when sugar prices fall below (say) 10 cents per pound, as they might be covering some of the fixed costs of production. In the long run, however, these firms will move their capital into uses that provide positive returns. Figure 8 below provides a basic outline of the way that these sugar and corn sweetener products enter their markets, and then brings them together via their use in the production of other commodities with the TRQ program on sugar-containing products..

(36) 34 Figure 8 FINAL USES. CES NEST (Armington). Sugar containing products dom. Sugar containing products imp. CRESH NEST Domestic sweetener using industries. Multi-source by TRQ allocation. CRESH NEST. C83 Refined Cane Sugar. C84 Refined Beet Sugar. C76 Corn Sweeteners. The U.S. sugar support program reaches down through the production chain into the final demand for sugar and other sweeteners. Imported products that contain sugar are analysed for sugar content, and then duties are applied on the basis of the sugar they contain. A TRQ program is used to administer these duties, with the intention of limiting the ability of final users of sugar containing products to implicitly substitute into ‘foreign sugar’ by sourcing products that contain it. This also provides some protection for the corn sweetener industry indirectly, as it also limits the downstream competition for their products. At the time of writing, this particular arm of the TRQ program is still being refined in MONASH-USA, mainly due to the limitations on data availability and the complexity involved in generating accurate duties for the various sugar-containing products under the MONASH-USA commodity classifications. The nesting of the imported products by region is identical in structure to that for sugar and corn sweeteners, and a final composite enters the model for downstream use..

(37) 35 The CRESH nest that combines the sweetener products was facilitated by adding an extra level of nesting to the MONASH-USA production structure. This is not an insignificant change, and is worth briefly discussing. Figure 9 below outlines the production structure of MONASH-USA and highlights the additional level of the intermediate goods nest:.

(38) 36 Figure 9. Output. Leontief nest Composite good g. Primary factor composite Other cost tickets CRESH nest of primary factors. CRESH nest of commodity i Composite labour. Commodity i. CES nest of source s Commodity i from source s. Land. Composite capital. CRESH nest of occupation m. CRESH nest of capital type n. Occupation m. Capital type n. CRESH nest of good g Commodity i from domestic source. Commodity i from imported source. Composite commodity g into creation of capital type n CRESH NEST Multi-source by region. CES nest of source s. Commodity i from source s into composite commodity i. Commodity i from domestic source. C81 Raw Cane Sugar imp CRESH NEST Multi-source by region.

(39) 37 The MONASH-USA production nest is a multi-level nested function of labour, capital, land, intermediate goods and “other cost tickets” (i.e. costs not explicitly modelled in MONASH-USA with behavioural functions - for example, the costs associated with holding inventories).. In. previous versions of MONASH-USA, the intermediate usage of goods was modelled as a threestage nest: imported commodities were CRESH nested by region; then an Armington (CES) nest of the domestic and foreign sources was applied, and thirdly, the resulting vector of composite intermediate goods was utilised in fixed proportions that are pre-determined by technology (effectively, a Leontief nest).. To allow for the alternative sweeteners to be substitutable in. production, MONASH-USA now applies a CRESH nest to the vector of intermediate good composites prior to entering the final level of the production function. As of the time of writing, the array of substitution elasticities in the database contains all zeros except for refined cane sugar, refined beet sugar and corn sweeteners. The non-zero elasticities are applied only in cases where all three are used by an industry. With the CRESH nest in place, it would be possible to fill this array of 262,636 cells with individual estimated elasticity values, but in most cases we would expect that these substitution elasticities would be quite small as technological and technique-related considerations will largely determine the input choice for most industries as regards intermediate commodity usage.. 4.3 A Generalised Complementarity Relationship This section is a brief summary of elements of chapter 16 in GEMPACK documentation, and readers are referred to this text or Harrison et al (2002) for more detail. It is presented to provide some background and to aid in the understanding of the complementarity relationships developed for MONASH-USA below. Complementarity relationships describe the values that a variable can take when some other condition – perhaps an expression – is observed to take one of a set of possible states. Denote the ‘complementarity variable’ by X, the ‘complementarity expression’ by EXP, and lower and upper bounds on X by L and U respectively. Mathematically, a generalised complementarity relationship takes the form. L <= X <= U ⊥ EXP , which is notation for the set of relationships. (2).

(40) 38 Either. X =L. and. EXP > 0. [state 1]. or. L < X <U. and. EXP = 0. [state 2]. or. X =U. and. EXP < 0 .. [state 3]. This three-state complementarity can also take the form of one of a pair of two-state relationships [which are really special cases of (2)] with either (a) a single lower bound, where Either. X =L. and. EXP > 0. [state 1]. or. X >L. and. EXP = 0. [state 2]. or (b) a single upper bound, where Either. X <U. and. EXP = 0. [state 2]. or. X =U. and. EXP < 0 .. [state 3].. These relationships can be summarised diagrammatically as follows:.

(41) 39 Figure 10 X. State 3 U. State 2. State 1 L EXP 0. The two-state complementarity relationships are represented on figure 10 by simply removing state 3 and U for the lower-bound-only case, and state 1 and L for the upper-bound-only case. Two of the applications below (for calculating the total value of quota rents and tariff revenue under the TRQ regime) utilise another characteristic of complementarities that is worth briefly describing here. Generally, a complementarity with a single bound (a dual-state complementarity) can be expressed as a MAX (maximum) or MIN (minimum) expression. A complementarity with only a lower bound can be stated as Either. X −L=0. and. EXP > 0. [state 1]. or. X −L>0. and. EXP = 0 ,. [state 2]. which is the same as MIN ( X − L, EXP ) = 0 ,. (3).

(42) 40 and for the upper bound case, by the same logic, we can reformulate and restate the relationship by Either. X − U <= 0. and. EXP = 0. [state 2]. or. X −U = 0. and. EXP < 0 .. [state 3]. and MAX ( X − L, EXP ) = 0 .. (4). In our application to the modelling of the U.S. sugar sector in MONASH-USA, all three complementarity types are utilised.. 4.4 The MONASH-USA Theory of the U.S. Sweetener Sector 4.4.1 Introduction The methodology described in this section is an extension of earlier work by Elbehri and Pearson (2000) and (particularly) Harrison et al (2002). The underlying method used for the modelling of complementarities also owes much to Ken Pearson, Mark Horridge and Jill Harrison, although the current paper mainly describes some extensions developed during the current project. In one sense, the differences between the former studies and the model described here are analogous to the differences between figures 1 and 5 above. One extension is applied mainly as a result of the U.S. domestic support program, and would be true in many other examples of tradable commodities related by production to agricultural crops. In this study, the rents generated under the U.S. policies are divided into two components – a “first tier” rent per unit related to the support price, and a “second tier” rent per unit generated by the quantity constraint at the quota boundary. An import diagram like figure 1, but slightly modified, can highlight these concepts..

(43) 41 Figure 11. $US. PUS3 = Pw(1+ti)(1+ φˆ 2 ) = Pw(1+to) PUS2 = Pw(1+ti)(1+ φ 1 )(1+ φ12 ). PS = Pw(1+ti)(1+ φ 1 ). PUS1 = Pw(1+ti) Pw. t. u. t. v. D3. s r o. q. p. k. l. m. f. g. h. n i. D2 j. D1 a. b M1. c M1’. d M2 = Q. e M3 Imports. In contrast to figure 1, the rents described in figure 11 involve three types of per-unit rent variable, denoted φ 1 , φ 2 and φˆ 2 , and a support price, denoted Ps. Variable φ 1 represents the rent on a unit of sugar due to the support price (or “first-tier” rent), φ 2 denotes a variable rent per-unit caused by the quota boundary (or “second-tier” rent), and φˆ 2 now represents the maximum value of. φ 2 specifically. The underlying theory of a TRQ captured in figure 11 differs slightly from that in figure 1. This is due to a number of sugar-industry specifics, and also some differences between the theoretical bases of a model like that detailed in figure 1 and the model described in this paper. Briefly, these differences stem from: 1. The existence of the loans scheme and support price; 2. the existence of the domestic marketing allotments; 3. the obligation to import imposed on the U.S. by its WTO commitments; 4. the multi-source commitments imbedded in the TRQ quota allocations;.


Figure 1

Figure 1

Figure 2

Figure 2

Figure 3

Figure 3

Figure 4

Figure 4

Figure 5 y$US

Figure 5

y$US p.23
Figure 6

Figure 6

Figure 7

Figure 7

Figure 8

Figure 8

Figure 9 Output

Figure 9

Output p.38
Figure 10

Figure 10

Figure 11

Figure 11

Figure 12 TARrOVQ TAR  r

Figure 12

TARrOVQ TAR r p.58
Figure 13

Figure 13